Biblios

Inspiration

As students and researchers, we often struggle to organize and extract insights from the flood of academic papers available online. Searching, clustering, and finding meaningful directions in existing literature is time-consuming and inefficient. We wanted to build something that makes research as intuitive as thinking itself — an AI-driven assistant that organizes at the speed of thought.

What it does

Biblios is a no-code research organization tool built entirely using AWS PartyRock. Users can simply paste a topic or upload papers, and Biblios:

Discovers relevant papers and content.

Analyzes them to summarize, visualize trends, and identify knowledge gaps.

Explores connections through an interactive clustering system and recommends new directions or related papers.

Essentially, Biblios acts as a smart, iterative research assistant — turning scattered information into structured, actionable insights.

How we built it

Biblios was built end-to-end using AWS PartyRock, leveraging its LLM capabilities and drag-and-drop components to prototype rapidly without writing a single line of code. We focused heavily on Amazon’s “Working Backwards” principle — defining the final user experience and press release before building anything. That mindset helped us map the user journey, from inputting a topic to exploring an integrated research ecosystem with visual trend detection.

Through PartyRock, we combined multiple AI components — from summarization and clustering to iterative prompt refinement — all orchestrated within a no-code environment.

Challenges we ran into

Working without code forced us to think in system-level design terms instead of implementation details. We had to translate technical logic into component workflows that PartyRock could execute naturally. We also had to balance AI creativity with reliability — ensuring that generated clusters, insights, and recommendations remained accurate and consistent.

Accomplishments that I'm proud of

Placed 2nd in the AWS PartyRock Hackathon, among several strong competitors.

Built a functional, no-code research insight engine in under 4 hours.

Developed a reusable design framework for other no-code AI workflows.

Learned to apply the Working Backwards philosophy effectively — designing first for impact, not features.

What I learned

This hackathon pushed me to think beyond coding — toward product thinking, business sense, and customer experience. By removing code from the equation, I learned how to articulate technical logic in simple, modular workflows, how to prioritize usability, and how to think like a product owner instead of just a developer.

Built With

• amazon-web-services
• partyrock